Participatory sensing is a revolutionary paradigm where people sense their environment, and voluntarily share this information using pre-existing communication infrastructure. It has tremendous potential to advance knowledge in diverse fields including urban planning and intelligent transportation by enabling large-scale sensing without dedicated infrastructure. Achieving this potential hinges on encouraging sensor data sharing between users who may not trust each other.
The objectives of this project are to design and implement a trustworthy participatory sensing system, encompassing algorithms for (i) certifying the integrity of published sensory content, so users can trust it, (ii) sensory content protection, so more users contribute data, and (iii) anonymous content sharing between producers and consumers. It explores a novel approach for trustworthy participatory sensing with a trusted hardware platform for each sensor, which besides the main processor, has a tamper-proof trusted hardware element that only executes code signed by a trusted third party. This recasts the integrity and privacy problems in a new way, allowing for simple, scalable and versatile solutions. Such a system can curb potential abuse, encourage user participation and thereby enable participatory sensing.
Expected project results include a new system model and algorithms for trustworthy participatory sensing, an implementation on the Intel AMT platform, and system demonstration in a real-world intelligent transportation application. The results will be disseminated via open-source software, conference publications and exhibits at the Oregon Museum of Science and Industry. The project also consists of multi-level educational activities to promote under-represented groups in computing.
Today, there are over 4 billion cellphone users in the world; many with smart phones. These smart phones, equipped with positioning, audio, video and other sensors, offer an unprecedented opportunity to instrument our world for applications ranging from intelligent transportation, urban monitoring, public safety to public health. "Participatory sensing" or "mobile crowdsourcing" is a revolutionary new paradigm where people voluntarily sense their environment using the built-in sensors in mobile phones and share this information using the existing infrastructure of the Internet and cellular networks. Using citizen contributed data for the above applications poses questions as to the integrity of the data. To understand how easy it is to launch software-based cyber attacks against such open systems, consider the following example. In June 2013, Google acquired Waze, a social navigation company for nearly $1.3 billion. Waze learns about traffic problems through crowdsourced road condition reports from users, reducing driver commutes. However, in March 2014, two Israeli students showed just how easy it was to launch cyber attacks on such a system. They registered thousands of fake Waze users, using a program that impersonated smartphones, and contributed false GPS coordinates to the app from these accounts (e.g. submitting reports claiming to be stuck in traffic at the false coordinates). As a result, the students created a virtual traffic jam. Technologies developed in this award make it impossible to launch such Sybil attacks. With the support of this National Science Foundation CAREER award, we have researched and developed computing solutions to ensure the integrity of user contributed data [Dua14a]. We developed a novel trusted platform module (TPM) based system (the first such system!) that addresses the problem of providing data integrity in participatory sensing. The key idea is to provide a trusted platform within each sensor device to attest the integrity of sensor readings. This localizes integrity checking to the device, rather than relying on corroboration, making the system not only simpler, but also resistant to collusion and data poisoning. A "burned- in" private key in the TPM prevents users from launching Sybils. Data collection from mobile phones also compromises citizen privacy by tracking people's behavior to an almost unprecedented degree. We have developed solutions to protect user data and preserve location privacy (while safeguarding data integrity!) [Dua11a], using interactive proof-protocols and enable selected sharing of information [Dua13b], using broadcast encryption. With collaborators in Australia, we have developed systems to explore how mobile phones can be harnessed to collect consumer price information and track market price dispersion [Bulusu08a]. While this work is fairly recent, it has captured the attention of government agencies! The Australian government wanted to set up a website called GroceryChoice to enable consumers to compare grocery prices from many supermarkets across the country. This project has failed. The supermarkets cited the classical response that it was technologically difficult to provide data. However, it is impossible to rule out the fact that the dominant supermarkets simply did not want to provide the data. In October 2010, the Australian Senate invited the research team to appear and testify at a hearing to explore causes of why the government initiative failed. The committee heard more about how participatory sensing may be used for tracking price dispersion in fuel and grocery prices. We have also developed participatory sensing applications for monitoring noise pollution called Ear-Phone [Rana14a], and investigated how user context classification can be used to ensure scientifically valid measurements. EarPhone costs significantly less than traditional urban noise pollution mapping systems, while providing live data. This work captured significant attention from media outlets, such as ABC Science News, MIT Technology Review, and Wired Magazine in 2013; and the attention of Canadian urban planners who want to try these technologies. References [Bulusu08a] Nirupama Bulusu, Chun Tung Chou, Salil Kanhere, Yifei Dong, Shitiz Sehgal, and David Sullivan, "Participatory Sensing in Commerce: Using Mobile Phones for Tracking Market Price Dispersion", In Proceedings of the International Workshop on Urban, Community, and Social Applications of Networked Sensing Systems (UrbanSense'08),Raleigh, North Carolina, November 2008. [Dua11a] Akshay Dua, Nirupama Bulusu and Wu-chang Feng, "Privacy-preserving Online Mixing of High Integrity Mobile Multi-User Data", In Proceedings of the 7th International Conference on Security and Privacy in Communication Networks (SecureComm '11), London, UK, September 2011. [Dua13b] Akshay Dua and Nirupama Bulusu, "Resource-aware Broadcast Encryption for Selective Sharing in Mobile Social Sensing", In Proceedings of the 8th IEEE International Conference on Intelligent Sensors, Sensor Networks and Information Processing (ISSNIP '13), Melbourne, Australia, April 2013. [Dua14a] Akshay Dua, Nirupama Bulusu, Wu-chang Feng, and Wen Hu, "Combating Software and Sybil Attacks to Data Integrity in Crowd-Sourced Mobile Embedded Systems", ACM Transactions on Embedded Computing Systems (TECS), September 2014. [Rana14a] Rajib Rana, Chun Tung Chou, Nirupama Bulusu, Salil Kanhere, and Wen Hu, "Ear-Phone: A Context-Aware Noise Mapping Using Smart Phones", Elsevier Pervasive and Mobile Computing, 2014.
